The Agent Skills Directory

[SAFE]: The Python scripts (security_scanner.py, vulnerability_assessor.py, and compliance_checker.py) are implemented using standard libraries and perform static analysis via regular expression matching. They are designed to read and report on file content without executing it or downloading external payloads.- [SAFE]: No data exfiltration or credential harvesting patterns were detected. The scripts operate exclusively on the local filesystem of the target directory provided by the user and do not initiate any network connections.- [SAFE]: The provided documentation and code examples in the references/ directory follow industry-recognized security best practices, such as the OWASP Top 10 and SOC 2/PCI-DSS standards. These materials are instructional and promote secure coding.- [INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for untrusted data as it reads external source code and dependency manifests. However, this does not pose a security risk as the skill lacks exploitable capabilities like dynamic code execution or network communication that could be triggered by malicious data. Evidence:
Ingestion points: Files are read in scripts/security_scanner.py, scripts/vulnerability_assessor.py, and scripts/compliance_checker.py using Path.read_text().
Boundary markers: Absent.
Capability inventory: Operations are limited to local file reads and printing analysis results; no subprocess, eval, or exec calls are performed on ingested content.
Sanitization: Content is processed solely through static regex filters.

senior-secops